EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-SA 4.0

Performance Evaluation of Regional Heatwave Prediction Using Statistical and Deep Learning Models

SULOCHANA DEVI, RADHIKA KOTECHA

Journal of Agrometeorology · 04.06.2026 · 28 (2) · 237-248

DOI: 10.54386/jam.v28i2.3400 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Agrometeorology, online erschienen am 04.06.2026. Verfasst von SULOCHANA DEVI und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Rising temperatures are increasing the frequency and impact of heat extreme events in India. So, this is increasing the need for early warning and climate-risk management. Heat risks and behaviors vary across regions, so national model can miss local temperature dynamics. We present a season-aware, region-wise framework to forecast daily maximum temperature and identify heatwave days in IMD defined Central India region. In Central India's severe pre-monsoon heat is recurrent and closely linked to crop-weather stress, irrigation demand, and outdoor labor exposure. Heatwaves are first categorized at grid-cell level using operational threshold-and -departure rules. Then heatwaves are aggregated to a region-day label using a spatial coverage threshold and minimum 2-day duration. In this paper, we benchmark conventional statistical time-series models against recurrent sequence models under same splits and evaluation window using national meteorological temperature data. Performance is evaluated with temperature-error metrics and event-based measures for heatwave-day detection. Results show that recurrent models give the strongest overall skill, with the LSTM model delivering the promising improvements. Seasonal statistical modeling improves over non-seasonal baselines by capturing the seasonal cycle. This framework offers a compact regional benchmark and practical guidance for region specific early warning systems. This work supports agrometeorological advisories for farm operations, water planning and heat-stress risk management during extreme heat.

Zitieren

APA

DEVI, S., & KOTECHA, R. (2026). Performance Evaluation of Regional Heatwave Prediction Using Statistical and Deep Learning Models. Journal of Agrometeorology, 28(2), 237-248. https://doi.org/10.54386/jam.v28i2.3400

Vancouver

DEVI S, KOTECHA R. Performance Evaluation of Regional Heatwave Prediction Using Statistical and Deep Learning Models. Journal of Agrometeorology. 2026;28(2):237-248. doi:10.54386/jam.v28i2.3400

BibTeX
@article{DEVI2026,
  author = {DEVI, SULOCHANA and KOTECHA, RADHIKA},
  title = {Performance Evaluation of Regional Heatwave Prediction Using Statistical and Deep Learning Models},
  journal = {Journal of Agrometeorology},
  year = {2026},
  volume = {28},
  number = {2},
  pages = {237--248},
  doi = {10.54386/jam.v28i2.3400}
}
Kostenlos starten →